civodul has quit [Quit: ERC (IRC client for Emacs 25.3.1)]
hakujin has quit [Ping timeout: 255 seconds]
BlessJah has quit [Ping timeout: 256 seconds]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to master: https://git.io/vAWmM
<NixOS_GitHub>
nixpkgs/master 2173402 Nick Novitski: reflex: init at 0.2.0
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 0b09136 Joachim F: Merge pull request #35053 from nicknovitski/reflex-init...
yegortimoshenko has quit [Remote host closed the connection]
kakuzade has quit [Quit: Page closed]
yegortimoshenko has joined #nixos
Fare has quit [Ping timeout: 248 seconds]
<symphorien>
that's strance
<symphorien>
*strange
hakujin has joined #nixos
<infinisil>
symphorien: the crate?
<symphorien>
that it ignores environment
<infinisil>
Indeed
<infinisil>
Goddamnit, why can i never develop in peace for once
<infinisil>
always trouble
kakuzade has joined #nixos
hakujin has quit [Ping timeout: 256 seconds]
kakuzade has quit [Client Quit]
BlessJah has joined #nixos
<bsima>
what does the `//` operator do, exactly?
<symphorien>
it merges two attrsets
<bsima>
ok
<bsima>
in `a // b` which will take precedence when there are duplicate attrs?
<symphorien>
the second one iirc
<bsima>
thanks symphorien
erasmas has quit [Quit: leaving]
pie___ has joined #nixos
jsgrant_ has quit [Quit: Leaving]
pie__ has quit [Ping timeout: 248 seconds]
EG^ has joined #nixos
hakujin has joined #nixos
<infinisil>
symphorien: Alright Ralith helped me out in mozilla's #rust, turns out the latest git version of the crate had the bindgen result included now, so I'm just using that now
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Drakonis has joined #nixos
aanderse_ has joined #nixos
aanderse has quit [Ping timeout: 240 seconds]
ryanartecona has quit [Quit: ryanartecona]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] nicknovitski opened pull request #35063: yarn: add yarnpkg binary to outputs and use nodejs dependency (master...yarnpkg) https://git.io/vAW3W
NixOS_GitHub has left #nixos [#nixos]
fragamus has joined #nixos
<infinisil>
You know what I need
dnovosel has quit [Remote host closed the connection]
<infinisil>
A module for each editor, and modules for each programming language. Enabling a language sets up each enabled editor to have a development setup for this language
<infinisil>
that would be nice
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
aanderse has joined #nixos
dnovosel has joined #nixos
aanderse_ has quit [Ping timeout: 255 seconds]
must^ has quit [Ping timeout: 260 seconds]
justanotheruser has joined #nixos
freeman42x]NixOS has quit [Ping timeout: 264 seconds]
vidbina has quit [Ping timeout: 256 seconds]
dan_b has quit [Ping timeout: 248 seconds]
ambro718 has quit [Ping timeout: 260 seconds]
aanderse_ has joined #nixos
armijn has quit [Remote host closed the connection]
aanderse has quit [Ping timeout: 264 seconds]
ottidmes has quit [Ping timeout: 240 seconds]
jb554 has quit [Ping timeout: 248 seconds]
<bsima>
does anyone know how I would modify a haskell build to cp some static files into $out? Adding the `cp` in the pre/postInstall phase doesn't seem to be working
<joepie91>
running nixos-install on a new system, with a large config
<joepie91>
runs out of space because it seems to install everything in /nix/store first, before copying it over to /mnt/nix/store?
<joepie91>
and the total size of all my stuff exceeds the size of the /nix/store mount (which is presumably in memory)...
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] rnhmjoj opened pull request #35064: nixos/networking: start dhcpcd even when defaultGateway6 is set (master...dgw) https://git.io/vAWZG
NixOS_GitHub has left #nixos [#nixos]
Fare has joined #nixos
hakujin has joined #nixos
sud0x3 has quit [Ping timeout: 256 seconds]
hakujin has quit [Client Quit]
silver has quit [Read error: Connection reset by peer]
<infinisil>
joepie91: you could use a very minimal config for the install, then boot into the system and do the full config
<joepie91>
infinisil: yeah, I was just hoping not to have to do that :/
<joepie91>
not sure I see why it needs to install everything into RAM first...
<infinisil>
joepie91: whatever /nix/store is for the currently running system, doesn't have to be in RAM i think
<joepie91>
it is on the live distribution
<infinisil>
the default nixos installer just is
<joepie91>
:p
<Acou_Bass>
joepie91: i had that issue... can comment out most of your config then rebuild when its installed, or create a swap partition/ file somewhere xD
<clever>
joepie91: its a bug in the current nixos-install
<clever>
joepie91: enabling swap can help, and you can resize the fs that has failed with mount /nix/.rw-store -o remount,size=3G
<joepie91>
I don't have swap :P
aanderse_ has joined #nixos
<clever>
joepie91: you can still resize it, to allow it to consume more ram
<clever>
it defaults to 50% of the ram
<joepie91>
right
<joepie91>
well either way, I installed a more minimal version now
<joepie91>
and as usual it fails on the bootloader setup :P
aanderse has quit [Ping timeout: 276 seconds]
<clever>
gpt or mbr?
<joepie91>
GPT, but I probably just set it up wrong
<clever>
efi or legacy?
<Fare>
gpt sounds weird in French (it pronounces like "I farted")
<clever>
lol
<joepie91>
clever: EFI, but I forgot to import the right config file
<infinisil>
joepie91: btw, you can just append -I nixos-config=path/to/nixpkgs/nixos/modules/profiles/minimal.nix for a minimal config
coot has quit [Quit: coot]
pie___ has quit [Ping timeout: 240 seconds]
<nhill>
So, I'm installing on a laptop with a Broadcom BCM43142 wireless card
<nhill>
I've put boot.extraModulePackages = [ config.boot.kernelPackages.broadcom_sta ];
<nhill>
in configuration.nix, but it does not load the package
<nhill>
lsmod confirms no kernel modules are loaded
<nhill>
what am I missing?
fragamus has joined #nixos
<infinisil>
nhill: did it auto-generate that with nixos-generate-config?
<joepie91>
aaaaaaaaaa
<nhill>
infinisil: no, it's not in the hardware file
<joepie91>
GRUB keeps complaining that my definitely-an-EFI-partition doesn't look like an EFI partition
<nhill>
I had to add it
<infinisil>
nhill: because for me it did, and it also put boot.kernelModules = [ "wl" ];
<nhill>
hmmm, I'll try that
<infinisil>
joepie91: did you set grub.device = "nodev"?
<joepie91>
infinisil: yep
<infinisil>
and efiSupport = true;?
<joepie91>
also manually mounted /boot which for some bizarre reason wasn't autodetected by generate-config either
<joepie91>
yeah
<infinisil>
nixos-generate-config doesn't mount anytihng
<joepie91>
sure, two separate issues
<joepie91>
it *was* autodetected on my desktop[
<joepie91>
hold on...
<clever>
joepie91: was /mnt/boot mounted before you ran nixos-generate-config?
<infinisil>
joepie91: maybe check `sudo blkid /dev/sd..` for your boot device
<infinisil>
should say efi something
d4g_ has joined #nixos
<joepie91>
clever: the problem seems to be that I mounted it in /boot rather than /mnt/boot
<joepie91>
(it was mounted in neither before)
<clever>
joepie91: ah, that will do it
<joepie91>
infinisil: already verified with gdisk that it really is an EFI partition
<joepie91>
there
<joepie91>
installation successful
<joepie91>
:P
<infinisil>
\o/
<joepie91>
let's give it a spin...
<joepie91>
meh, now my Windows boot entry is gone
d4g has quit [Ping timeout: 255 seconds]
<infinisil>
joepie91: you can give boot.loader.grub.useOSProber a try
<joepie91>
seems I can still boot Windows from the manual boot selection menu
<joepie91>
ah
vidbina has quit [Ping timeout: 248 seconds]
DavidEGrayson has joined #nixos
<DavidEGrayson>
Yo, is there a good way to tell if a given Nix derivation has been build yet? Checking for the existence of the output is not good enough because maybe it is in the middle of being built, or the system crashed while it was being built.
<DavidEGrayson>
So, given a .drv file path or the output listed in that, can I tell if it was built yet?
<joepie91>
infinisil: giving that a spin now and installing my full set of applications too :)
<clever>
DavidEGrayson: hmmm, one min
<clever>
DavidEGrayson: i think you can run `nix-store --query --size` on the output
mbrgm has quit [Ping timeout: 268 seconds]
<clever>
DavidEGrayson: unfinished builds should come up as invalid
aanderse has joined #nixos
mbrgm has joined #nixos
aanderse_ has quit [Ping timeout: 256 seconds]
<joepie91>
man, SSDs make a massive difference
<samueldr>
I concur
ertes has joined #nixos
Lisanna_ has joined #nixos
lopsided98 has quit [Remote host closed the connection]
<joepie91>
rebuild complete...
<joepie91>
reboot!
lopsided98 has joined #nixos
<joepie91>
hahaha, Windows cheaped out with the boot animation
<joepie91>
it overlays the spinner on the NixOS GRUB
<clever>
lol
<joepie91>
they don't blank the screen first...
<joepie91>
I bet that's something to do with vendor branding actually
<joepie91>
anyway, booting into windows works
<joepie91>
thanks infinisil
<joepie91>
... now my graphical environment on NixOS is broken? what?
<joepie91>
well this is bizarre
<joepie91>
my graphical environment was running just fine
<joepie91>
but for some reason it got stuck on tty 1
<joepie91>
and I had to manually ctrl+alt+f7 to see it..
<DavidEGrayson>
clever: I think that works, thanks!
Supersonic has quit [Disconnected by services]
aanderse has quit [Ping timeout: 248 seconds]
Supersonic112 has joined #nixos
Supersonic112 is now known as Supersonic
ryanartecona has joined #nixos
<DavidEGrayson>
However, I do want to query hundreds of derivations at a time. Maybe I'll make a pull request to add 'nix-store --query --built' or something.
<DavidEGrayson>
Because --size stops after the first error.
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Drakonis has quit [Read error: Connection reset by peer]
<DavidEGrayson>
I am also tempted to just use an SQLite binding to open the Nix store database directly (and will accept it if I shoot myself in the foot).
<clever>
DavidEGrayson: it would be better to use the nix bindings
<clever>
DavidEGrayson: one min
<clever>
if (store->isValidPath(expectedStorePath))
<clever>
DavidEGrayson: this function tells you if a path is valid
<nhill>
So, how come users can't manage their own packages declaratively?
pkill9 has quit [Ping timeout: 260 seconds]
<DavidEGrayson>
I think it's easy to access SQLite databases from Ruby, and the schema of the Nix is so simple and easy to understand. I came up with this query to check the validity of each output of a list of drv files:
<DavidEGrayson>
select d.path, o.id, v.id from ValidPaths d left join DerivationOutputs o on d.id == o.drv left join ValidPaths v on o.path == v.path where d.path in ("/nix/store/foo.drv", ...);
<DavidEGrayson>
So that's pretty cool. It gives the ID of each output that is a valid path. For outputs that aren't a valid path, the ID is null.
ryanartecona has joined #nixos
iyzsong has joined #nixos
ryanartecona has quit [Quit: ryanartecona]
Ubercow has quit [Remote host closed the connection]
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<taktoa>
nhill: you can, read about ~/.config/nixpkgs/config.nix in the Nix manual
jrolfs has quit [Ping timeout: 268 seconds]
<clever>
DavidEGrayson: that also works
<nhill>
taktoa: took a while, realized its in the nixpkgs manual not nixos
Svarog has joined #nixos
<Svarog>
does the default firewall configuration on nixos block ports even for localhost? and if so how would i go about opening some ports so that i can access them locally?
nuncanada has quit [Remote host closed the connection]
nuncanada has joined #nixos
davidak has quit [Remote host closed the connection]
davidak has joined #nixos
Arcaelyx_ has joined #nixos
Arcaelyx has quit [Ping timeout: 252 seconds]
davidak has quit [Ping timeout: 256 seconds]
schoppenhauer has quit [Ping timeout: 265 seconds]
<gchristensen>
anyone here have bluetooth headphones working?
schoppenhauer has joined #nixos
ryanartecona has joined #nixos
<tertleeltret>
someone asked that yesterday in #funtoo lol
Lisanna_ has quit [Quit: Lisanna_]
<gchristensen>
I don't actually have any, but am considering some
<eacameron>
How can I see the volume controls in NixOS?
<eacameron>
I can't adjust the volume anywhere
<gchristensen>
how about nix-shell -p pavucontrol --run pavucontrol
marusich has joined #nixos
<eacameron>
gchristensen: Thanks. However, the window just hangs showing "Establishing connection to PulseAudio. Please wait..."
<gchristensen>
How about alsamixer
<eacameron>
gchristensen: Hahaha yeah that appears to be working.
<eacameron>
gchristensen: Kinda low-tech tho
<gchristensen>
Well now you can look up alsa mixers which suit your preferences
jsgrant has quit [Remote host closed the connection]
fragamus has joined #nixos
<eacameron>
grahamc: Ah...sadly I'm not sure how to do that?
<Olgierd>
hi, how hard would it be to start using gnome shell on wayland?
Rusty1_ has quit [Quit: Konversation terminated!]
Fare has quit [Ping timeout: 240 seconds]
Guanin has quit [Ping timeout: 240 seconds]
<sphalerite_>
gchristensen: I have a bluetooth speaker working, I imagine headphones would work pretty much the same. The one issue I have with bluetooth devices in general is that bluetooth and wifi interfere with each other and I get abysmal download speeds (in the tens of kB/s) when using any bluetooth devices
yegortimoshenko has quit [Remote host closed the connection]
yegortimoshenko has joined #nixos
nflores has joined #nixos
<Olgierd>
and for how much time is a nixos release supported? I can't find that on the website
<sphalerite_>
Olgierd: until the next one comes out
Guanin has joined #nixos
<sphalerite_>
Olgierd: so ~6 months
<sphalerite_>
Olgierd: we don't have the resources for LTS
iyzsong has quit [Ping timeout: 260 seconds]
pkill9 has joined #nixos
<Olgierd>
ty
hamishmack has quit [Quit: hamishmack]
pkill9 has quit [Ping timeout: 256 seconds]
endformationage has quit [Quit: WeeChat 1.9.1]
<mkaito>
yo nix wizards. say I've got a bash script, which I would like to deploy via nixops to a server, along with a systemd timer/service pair to run it periodically. what is the correct way to go about this?
<sphalerite_>
mkaito: something like systemd.timers.foo = {timer config}; systemd.services.foo = {script = ''script goes here''; path = with pkgs; [dependencies of foo]};
<nflores>
I just noticed that fetchurl has the ability to pass in curl flags using NIX_CURL_FLAGS: https://github.com/NixOS/nixpkgs/blob/master/pkgs/build-support/fetchurl/default.nix#L30. How do you set this variable? I was trying just NIX_CURL_FLGAS="-J -L -H "Accept: application/octet-stream"" within the same fetchurl block but doesn't quite work.
<sphalerite_>
nflores: impureEnvVars are propagated from the environment that started the build, not within nix
<sphalerite_>
nflores: so you set them at the point where you run nix-build
<sphalerite_>
hyper_ch: it does appear to be in the channel to me…
<leons>
I'm trying to build a rust application (not packaging for nixos, just with cargo) and want to speak to a sqlite database.
<leons>
However, the linker can't find the library: "ld: cannot find -lsqlite3"
<leons>
I'm not quite sure whether this is the right place to ask but does anyone know, which packages could resolve the issue? I've tried installing "sqlite" but that didn't help.
<NixOS_GitHub>
nixpkgs/gcc-7 7a1462c Vladimír Čunát: Merge branch 'master' into gcc-7...
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/gcc-7 d74d79a Vladimír Čunát: qt4: fix evaluation on Darwin after parent commit
<NixOS_GitHub>
[nixpkgs] vcunat pushed 2 new commits to gcc-7: https://git.io/vAWzK
<vidbina>
Does anyone know if overrideAttrs should work on texlive packages? The code is not failing, but I'm also not noticing a rebuild of the respective package. Perhaps there is a better way to modify texlive packages?
ma271 has joined #nixos
alex`` has quit [Quit: WeeChat 2.0.1]
<hyper_ch>
clever: that does not work
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] zimbatm opened pull request #35068: tmpdir audit: only fail with files referenced below (master...stricted-tmpdir-audit) https://git.io/vAWzF
<ambro718>
I am unable to get nixops to deploy a virtualbox vm (new one), the VM starts but it is not possible to ssh to it (or ping it). I enabled the DHCP server in virtualbox, did not help.
<ambro718>
How do I even log in to the machine in the console? I don't know the password.
JayVii has quit [Ping timeout: 264 seconds]
nico202 has joined #nixos
freeman42x]NixOS has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm closed pull request #24664: nginx: Check config syntax validity with `nginx -t` in preStart (master...nginx-check-config) https://git.io/vSg78
NixOS_GitHub has left #nixos [#nixos]
<nico202>
Hi, is there some nix hackery to download the sources of a package?
<joachifm>
nico202: as in nix-prefetch-url -A foo.src?
<Dezgeg>
nix-build -A foo.src
<nico202>
Dezgeg: yes! thanks
<nico202>
joachifm: yeah also this works, thanks
<Dezgeg>
nix-prefetch-url will probably not work for fetchgit and some others, I think
<joachifm>
true, it expects an urls attr, so probably not
ottidmes has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm closed pull request #34776: figlet: simplify, enable tests, fix for cross and musl (master...fix/fidget-cross) https://git.io/vAIh8
taktoa has quit [Remote host closed the connection]
ma272 has quit [Ping timeout: 260 seconds]
ma272 has joined #nixos
taktoa has joined #nixos
ThatDocsLady has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
alex`` has joined #nixos
alex`` is now known as alexherbo2
alexherbo2 is now known as alex``
orivej has joined #nixos
simendsjo has quit [Ping timeout: 256 seconds]
iyzsong has quit [Ping timeout: 265 seconds]
<hyper_ch>
rss2email is really annoying... I can just sendmail user@remote from the cli and it works all find... but r2e causes senmail exited with code 1
<hyper_ch>
wtf
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to master: https://git.io/vAWVC
<NixOS_GitHub>
nixpkgs/master 2e45234 Rick Yang: softether: 4.20 -> 4.25
<NixOS_GitHub>
nixpkgs/master 828b635 Joachim F: Merge pull request #34997 from rick68/softether...
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 1 new commit to master: https://git.io/vAWVl
<NixOS_GitHub>
nixpkgs/master 635011f Luke Sandell: wsjtx: init at 1.8.0 (#33247)
NixOS_GitHub has left #nixos [#nixos]
<joachifm>
Mic92: d'oh I see
ThatDocsLady has quit [Remote host closed the connection]
davidcl has quit [Quit: Ping timeout (120 seconds)]
asuryawanshi has quit [Ping timeout: 256 seconds]
<joachifm>
NinjaTrappeur: it's my fault ... i'm working on fixing it
davidcl has joined #nixos
<NinjaTrappeur>
Oh, ok, no problem!
<LnL>
we can trigger an eval when master is back to normal
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] NickHu opened pull request #35069: tcl/tk: 8.6.6 -> 8.6.8 and create library symlink (master...tcl) https://git.io/vAWVS
NixOS_GitHub has left #nixos [#nixos]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 1 new commit to master: https://git.io/vAWV9
<NixOS_GitHub>
nixpkgs/master 05f5cdc Joachim Fasting: Revert "Merge pull request #30031 from dtzWill/update/leo-5.6"...
NixOS_GitHub has left #nixos [#nixos]
jrolfs has quit [Ping timeout: 256 seconds]
<joachifm>
should be ok now ...
spear2 has quit [Quit: Leaving]
ThatDocsLady has quit [Ping timeout: 240 seconds]
NickHu_ has joined #nixos
JayVii has joined #nixos
FRidh has joined #nixos
thc202 has joined #nixos
<LnL>
NinjaTrappeur: there we go :)
<NinjaTrappeur>
Great, thanks :)
simukis has quit [Quit: simukis]
JayVii has left #nixos [#nixos]
<jluttine>
hmm.. one derivation sets preStart. how can i prepend something to this in my configuration.nix? if i just define something it is appended, if i understand correctly
pie___ has joined #nixos
Mateon3 has joined #nixos
<LnL>
mkBefore
<jluttine>
something like: foobar.preStart = mkPrepend "i want this in the beginning"
<jluttine>
LnL: ah!
<LnL>
and similarly mkAfter for the opposite
<NinjaTrappeur>
LnL, thanks for the tip regarding runCommand, that's really handy!
jrolfs has joined #nixos
Mateon1 has quit [Remote host closed the connection]
blahdodo has quit [Ping timeout: 256 seconds]
colescott has quit [Ping timeout: 256 seconds]
mr_sm1th has quit [Quit: No Ping reply in 180 seconds.]
Mateon3 is now known as Mateon1
mr_sm1th has joined #nixos
blahdodo has joined #nixos
hodapp has quit [Ping timeout: 256 seconds]
Fare has joined #nixos
hodapp has joined #nixos
<jluttine>
LnL: thanks! though, apparently it didn't solve my problem.. should be able to put the string to the middle..
<jluttine>
:(
<hyper_ch>
awwww. palemon compiling just takes like forever
<sphalerite_>
Anyone know how I can get a gdb that supports more arches?
<sphalerite_>
hyper_ch: yes, browsers and web rendering engines tend to be like that…
<sphalerite_>
hyper_ch: except dillo. Dillo compiles real quick
<hyper_ch>
I should go back to unstable
mojjo has joined #nixos
<mojjo>
hi!.. why does the following not work: toFile "ss" (toJSON (pkgs.writeShellScriptBin "myscript" "echo 123"))
<hyper_ch>
"Since the end of April 2015 and version 2015.04.26, youtube-dl uses -f bestvideo+bestaudio/best as the default format selection (see #5447, #5456)."
zzamboni has joined #nixos
<sphalerite_>
oh neat
<hyper_ch>
question is, do we use a newer version or not :)
habnabit has joined #nixos
xorkle_ has joined #nixos
<hyper_ch>
youtube-dl --version
<hyper_ch>
2018.02.08
<hyper_ch>
well, i do
bitchecker_ has joined #nixos
<joachifm>
mojjo: there's also unsafeDiscardOutputDependency which may allow you to work around the restriction you encountered ... can't speak to the sanity of that, however, depends on what you're doing ...
hodapp has quit [*.net *.split]
taktoa has quit [*.net *.split]
justanotheruser has quit [*.net *.split]
tmplt_ has quit [*.net *.split]
Xorkle has quit [*.net *.split]
akaWolf has quit [*.net *.split]
pmeunier has quit [*.net *.split]
bitchecker has quit [*.net *.split]
glines has quit [*.net *.split]
sweater has quit [*.net *.split]
fxkr has quit [*.net *.split]
octe has quit [*.net *.split]
_habnabit has quit [*.net *.split]
marcinkuzminski has quit [*.net *.split]
ncl28 has quit [*.net *.split]
statusbot has quit [*.net *.split]
joshie has quit [*.net *.split]
<joachifm>
mojjio: i'm guessing it's not what you want but still :)
<sphalerite_>
hyper_ch: you probably also want to actually use the youtube-dl input… do ${youtube-dl}/bin/youtube-dl instead of just youtube-dl otherwise it relies on youtube-dl being in the PATH
akaWolf has joined #nixos
<sphalerite_>
hyper_ch: you're also not using stdenv by name in there so you can remove that from the args
<hyper_ch>
youtube-dl is in path
ncl28 has joined #nixos
sweater has joined #nixos
<hyper_ch>
why shouldn't it be in path?
glines has joined #nixos
nico202 has quit [Ping timeout: 256 seconds]
ssmike has quit [Quit: ssmike]
<sphalerite_>
hyper_ch: if you've only installed your script but not youtube-dl
<hyper_ch>
sphalerite_: you're right
dredozubov has quit [Ping timeout: 260 seconds]
<hyper_ch>
already fixed it
pie___ has quit [Ping timeout: 256 seconds]
<sphalerite_>
err I'm stupid and can't read, you *are* using stdenv by name there
<sphalerite_>
but you don't need to if you use writeShellScriptBin instead :D
<hyper_ch>
why not?
marcinkuzminski has joined #nixos
dredozubov has joined #nixos
zzamboni has quit [Quit: Leaving.]
<sphalerite_>
because that adds the shebang for you
<Profpatsch>
*HUGS*
<hyper_ch>
sphalerite_: I don't know what you mean
<sphalerite_>
hyper_ch: that will add the #!${stdenv.shell} for you, and run a syntax check at build time :)
<hyper_ch>
sphalerite_: why would I wanna use that?
<sphalerite_>
because it's less verbose and has a syntax check
<hyper_ch>
less verbose than?
octe has joined #nixos
<sphalerite_>
writeScriptBin and putting the #! in yourself
<infinisil>
sphalerite_: admittedly there isn't much you can mess up with a single shell line haha
<hyper_ch>
sphalerite_: I don't see why that would be any beneficial
justanotheruser has joined #nixos
<sphalerite_>
infinisil: sure, but the tools are there sos why not use them
<infinisil>
ya
<sphalerite_>
hyper_ch: because #!${stdenv.shell} is avoidable boilerplate
<hyper_ch>
and that means?
<sphalerite_>
it's ugly and not useful
fxkr has joined #nixos
<sphalerite_>
Anyway, enough. I've probably wasted more time trying to convince you of this than you will ever waste writing #!${stdenv.shell} :p
<ottidmes>
infinisil: I am getting the help message of znapzendzetup during znapzend-pre-start. I copied over your service from master, and lib/types.nix, since you used the ints type, might I be missing something else? I checked the znapzend package itself, but I have the same as is on master, any idea?
etu has quit [Read error: Connection reset by peer]
<ottidmes>
infinisil: I just inspected some more and I guess the output of list is not right, causing your grep to return wrong info, which is then passed to znapzendzetup via xargs causing the help message
<sphalerite_>
Fare: or check the source to see if any special overriding is happening and if not nix-instantiate '<nixpkgs>' -A stdenv.cc
<ottidmes>
infinisil: The tilde is from my zsh showing the current directory, so you can ignore that bit, I should have removed it
simukis has joined #nixos
<infinisil>
ottidmes: does the same happen when you do systemctl restart znapzend
nico202 has joined #nixos
<Fare>
no overriding of cc happening, and regular gcc 6.4.0 wrapper is used. Weird.
<troydm>
guys I'm trying to build nixos package (kernel module) but it's built against a different version of kernel
<troydm>
how can I specify in nix-shell which kernel version to build against
<sphalerite_>
troydm: linuxPackages_x_y.<name>
<sphalerite_>
I'm trying to run some nix-built stuff on my VoCore, but am getting Illegal Instruction. The host system is openwrt and as such doesn't have a grand selection of debugging tools available. Does anyone know a way I can work out what's going on?
<ottidmes>
infinisil: Whohooh! Green light from systemctl status, thanks for the patch!
<infinisil>
ottidmes: \o/
<infinisil>
ottidmes: I shall make a PR to fix that right now
<infinisil>
Wait no, 2% battery remaining, I'll do it when I have power again heh
<troydm>
sphalerite_: ahh, silly me, thx
<genesis>
1,9G /tmp/nix-build-sdlmame-0.194.drv-0/ so after 4Hours of compilation i get a "No space left on device" :O
<genesis>
should try TMPDIR this time ^^
<genesis>
is there a way to resume the build from a nix-env -i ?
<sphalerite_>
Oh yes, I know the frustration… No space left on device, or out of memory, etc…
<genesis>
haha 3 weeks i try to package this stuff :3 should be easier with a good machine :)
<sphalerite_>
genesis: not really. You could do some really nasty stuff to cheat it but you wouldn't be able to tell if the build really works at that point
<sphalerite_>
for that you also need to have the build files lying around, which you don't unless you used --keep-failed
vcunat has joined #nixos
<genesis>
i always use -K yes :)
<sphalerite_>
in that case *maybe* you can start the build again, kill the make process with -STOP as root, then rsync in the old build files and see if that works. It's very ugly though, and not reproducible, and generally not advisable
<sphalerite_>
I'd use it to make sure a very expensive build works at least roughly, but not submit to nixpkgs without having done a clean build of it
<sphalerite_>
you may also need to use a debugger to have the make process re-exec itself, depending on how make works internally (I don't know the details)
<genesis>
yup
sud0x3 has quit [Ping timeout: 256 seconds]
kelleyNif has joined #nixos
sud0x3 has joined #nixos
<troydm>
sphalerite_: so, I have strange problem, I've specified right version of kernel for the module, and it built it against 4.14.9-dev, I'm running on 4.14.8. now when I try to insmod that built module it says Invalid module format
<troydm>
sphalerite_: I've built from locally cloned nixpkgs repo on 17.09 branch, and I'm running 17.09 myself
<troydm>
sphalerite_: I'm not sure why it chose 4.14.9 instead of 4.14.8, how can i check this?
sud0x3 has quit [Client Quit]
<troydm>
sphalerite_: also I'm not sure if those two issues are related tho
<sphalerite_>
troydm: yes, they are related. Modules have a specific version they're built against in their header or something, and the kernel will refuse to load them if the versions don't match
<troydm>
sphalerite_: so now the question is why did it chose 4.14.9 when stable channel is on 4.14.8
<sphalerite_>
troydm: how exactly did you build it? Do you perhaps have a more recent kernel in your system profile that you haven't booted yet? (compare readlink /run/{current,booted}-system/kernel)
<sphalerite_>
AFAICT it should be 4.14.18, not 4.14.8 or 4.14.9
<genesis>
nix-env seems to ignore my TMPDIR :/ , i tried both export and prepended
<troydm>
sphalerite_: both symlinks point to 4.14.8
<sphalerite_>
genesis: it's the daemon that spawns the build, so you'd have to change it for the daemon
kitemikaze has quit [Remote host closed the connection]
kitemikaze has joined #nixos
goibhniu has quit [Ping timeout: 260 seconds]
<genesis>
sphalerite_ : oki, i've to fix it in the systemd files i suppose.
humanoyd has joined #nixos
<sphalerite_>
yeah something like systemd.services.nix-daemon.environment.TMPDIR = "/path/to/alternative/dir"
nico202 has quit [Quit: Leaving]
<Fare>
what could cause two machines with identical nixpkgs checkout to have different cc -v ?
leat has quit [Quit: WeeChat 1.9.1]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/staging 0586da4 Frederik Rietdijk: Merge remote-tracking branch 'upstream/master' into HEAD
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to staging: https://git.io/vAWKM
NixOS_GitHub has left #nixos [#nixos]
<troydm>
sphalerite_: apparently kernel version is 4.14.9 for my locally git clone nixpkgs, also I'm on release-17.09 branch on it
jacob_ has joined #nixos
<troydm>
sphalerite_: where do I check this?
leat has joined #nixos
jacob_ is now known as Guest52362
kelleyNif has quit [Quit: Leaving.]
<sphalerite_>
troydm: you shouldn't generally use release-17.09, as it's not tested. Use nixos-17.09 from github.com/nixos/nixpkgs-channels instead. Is the repo actually up to date?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 2 new commits to master: https://git.io/vAWKF
<NixOS_GitHub>
nixpkgs/master de5bbd0 Linus Heckemann: firefox: enable pulseaudio by default...
<i-am-the-slime>
I got xmonad somewhat running, but it doesn't automatically connect to wifi (as plasma did). What are the options to get this working? I guess I could connect through the command line.
kitemikaze has quit [Remote host closed the connection]
<srhb>
i-am-the-slime: networkmanager is fairly common
<sphalerite_>
i-am-the-slime: I'm not sure but I think plasma pulls in network-manager. If you enable it it should use the settings you already have
<i-am-the-slime>
So I would install that?
<srhb>
i-am-the-slime: networking.networkmanager.enable = true; and you'll probably want to install the tray widget thingy
<i-am-the-slime>
Where would I put some kind of autostart for nm-applet?
<sphalerite_>
Fare: nix-shell -I nixpkgs=$nixpkgs -p --run 'cc -v' should produce the same for both machines as long as they're the same arch
tgunb has joined #nixos
<Fare>
ok, so actually ~/.nix-profile/bin/cc is the same on both machines
robstr has joined #nixos
<Fare>
wait, not anymore (!)
mkoenig_ has joined #nixos
<Fare>
cc -v returns the same on both machines, but your command returns something different. I'm so confused.
mkoenig has quit [Ping timeout: 276 seconds]
<sphalerite_>
Fare: again, compilers, interpreters, libraries and generally those sorts of tools shouldn't be installed into user profiles but used via nix-shell
<i-am-the-slime>
xmonad also ignores all my stuff about remapping keys in configuration.nix
<sphalerite_>
Fare: a compiler installed into your profile will also not find libraries for instance
<Fare>
sphalerite_, well, I'm doing interactive compilation at the REPL, so I pretty much need them.
<Fare>
and I extract configuration variables from nix-shell
<sphalerite_>
what's wrong with just using nix-shell?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Mic92 pushed 1 new commit to release-17.09: https://git.io/vAW69
<NixOS_GitHub>
nixpkgs/release-17.09 c882c4d Linus Heckemann: firefox: enable pulseaudio by default...
NixOS_GitHub has left #nixos [#nixos]
<hyper_ch>
sphalerite_: you made the zanpzend module right?
<sphalerite_>
hyper_ch: I think you might be confusing me with infinisil
<Fare>
why does the nix-shell have different paths on the two machines???
<hyper_ch>
sphalerite_: quite possible... you both sound the same
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to gcc-7: https://git.io/vAW6b
<NixOS_GitHub>
nixpkgs/gcc-7 1274db6 Vladimír Čunát: webkitgtk: build with gcc6 for now...
NixOS_GitHub has left #nixos [#nixos]
<Fare>
this is the same $nixpkgs, but a different checkout and configuration of nixos underneath
<Fare>
different kernels. Does that count?
tmplt__ has quit [Remote host closed the connection]
tmplt has quit [Remote host closed the connection]
<srhb>
Fare: Not if you're really doing '<nixpkgs>' -- that does not rely on anything NixOS.
<srhb>
Fare: Something else must differ.
<Fare>
I'm utterly baffled. And variadic functions don't work on gerbil on the $newlaptop, as well as plenty of discrepancy, despite having identical $nixpkgs at the base
tgunb has quit [Ping timeout: 252 seconds]
<srhb>
Fare: Inspect the graph with nix-store
vcunat has quit [Ping timeout: 260 seconds]
<troydm>
sphalerite_: well I've managed to built the module but now when I try to insert it it says Unknown symbol reference, because it references other module, any way to provide those references or how do I install my customly built kernel module?
robstr has quit [Remote host closed the connection]
robstr has joined #nixos
<sphalerite_>
Fare: nix-copy-closure --to <other-host> $(nix-instantiate $nixpkgs -A gerbil) and then run nix-diff against the two
vcunat has joined #nixos
<sphalerite_>
troydm: boot.extraModulePackages
<sphalerite_>
troydm: so something like boot.extraModulePackages = with config.boot.kernelPackages ; [ rtl8812au ];
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh created python-unstable (+21 new commits): https://git.io/vAWir
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/python-unstable 43f6eed Frederik Rietdijk: python: agate-excel: 0.2.1 -> 0.2.2
<NixOS_GitHub>
nixpkgs/python-unstable 826b7ad Frederik Rietdijk: python: alembic: 0.9.6 -> 0.9.8
<NixOS_GitHub>
nixpkgs/python-unstable adc5004 Frederik Rietdijk: python: agate-sql: 0.5.2 -> 0.5.3
<NixOS_GitHub>
[nixpkgs] oxij opened pull request #35071: stdenv, bash: fixing info pages and stuff (staging...stdenv/infopages) https://git.io/vAWii
NixOS_GitHub has left #nixos [#nixos]
Ross has joined #nixos
srhb has joined #nixos
vcunat has quit [Ping timeout: 260 seconds]
<troydm>
sphalerite_: I know how to specify them in configuration.nix, what I'm actually trying to do is this: 1) build custom wl.ko module with debug enabled (done) 2) remove system wl module (done) 3) insert custom using insmod wl.ko (issues)
<troydm>
sphalerite_: now my problem is that wl.ko references symbols from cfg802 module
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 3 new commits to python-unstable: https://git.io/vAWiS
<NixOS_GitHub>
nixpkgs/python-unstable 2328af3 Frederik Rietdijk: python: nbstripout: 0.3.0 -> 0.3.1
<NixOS_GitHub>
nixpkgs/python-unstable 03bd7de Frederik Rietdijk: python: khal: 0.9.7 -> 0.9.8
<NixOS_GitHub>
nixpkgs/python-unstable 867e26c Frederik Rietdijk: python: cheat: 2.2.1 -> 2.2.3
NixOS_GitHub has left #nixos [#nixos]
<troydm>
sphalerite_: apparently module symbols are specified in Module.symvers
<sphalerite_>
don't know about that stuff, sorry
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] FRidh pushed 1 new commit to python-unstable: https://git.io/vAWi7
<NixOS_GitHub>
nixpkgs/python-unstable 18b41d6 Frederik Rietdijk: python: guessit: 2.0.4 -> 2.0.5
NixOS_GitHub has left #nixos [#nixos]
Myrl-saki has quit [Quit: WeeChat 1.9.1]
<Fare>
how do I reset the entire installation of $nixpkgs and garbage collect it away?
<srk>
aminechikhaoui: it's just a jobsets = pkgs.writeText "spec.json" (builtins.toJSON jobsetsAttrs);
<srk>
maybe mandatoryFeatures=["local"]
smichel17 has joined #nixos
coot has joined #nixos
<srk>
nah, still stuck, maybe this combination is not good Hydra 2017-11-21 (using nix-2.0pre5889_c287d731).
seanparsons has quit [Quit: ZNC 1.6.5 - http://znc.in]
Ross has quit [Read error: Connection reset by peer]
<Fare>
why would the nix-shell's $PATH differ on the two machines with identical $nixpkgs ?
Neo-- has joined #nixos
newbie__ has joined #nixos
Ross has joined #nixos
<newbie__>
sorry for the newbie question. I ran nix-shell -p armadillo eigen , and then lauched cmake for a project using those libraries. They both were not found. Am I missing some steps? :/
<Fare>
OK, so the NIX_PATH is wrong on one machine
NickHu_ has quit [Ping timeout: 260 seconds]
seanparsons has joined #nixos
<i-am-the-slime>
kim0 has joined #nixos
i-am-the-slime has left #nixos ["Konversation terminated!"]
Myrl-saki has joined #nixos
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to gcc-7: https://git.io/vAWX9
<NixOS_GitHub>
nixpkgs/gcc-7 b31d598 Vladimír Čunát: gfortran: use 7 by default, so we match gcc
NixOS_GitHub has left #nixos [#nixos]
<vcunat>
newbie__: also add cmake on the -p list
<vcunat>
though I'm not certain it will really help
<newbie__>
trying right away :)
<newbie__>
well it did for armadillo :)
<vcunat>
it's about a hook that sets up env variables
<vcunat>
when addind build inputs
<newbie__>
now eigen, hmmm
<hyper_ch>
infinisil: got questions about znazpzend
<sphalerite_>
!doesanyoneuse
<{^_^}>
Don't ask if anyone uses software xyz, just ask your specific question. This helps solve your problem quicker, and allows people who haven't used xyz but may still be able to help you to do so.
<sphalerite_>
hyper_ch: ^
<hyper_ch>
sphalerite_: I didn't ask any question ;)
<hyper_ch>
sphalerite_: I was letting him just know ;)
<hyper_ch>
big difference
Ross has quit [Read error: Connection reset by peer]
<sphalerite_>
ok let me pretend to be infinisil. "ok, what are your questions?"
<hyper_ch>
but you are not as you've clearly stated before
<ottidmes>
hyper_ch: I just made znapzend work on my server
<hyper_ch>
ottidmes: I can't run znapzend due to encryption
alex`` has joined #nixos
<infinisil>
haha
<hyper_ch>
there he is
<hyper_ch>
he pretended like he doesn't use the computer on weekends it seems :)
<infinisil>
i now get notifications on my phone when somebody mentions me :)
<hyper_ch>
the question I had was already answered :(
<hyper_ch>
sorry for making your phone notify you
<hyper_ch>
znapzend has a daemon mode :(
bluemonk12 has joined #nixos
<infinisil>
what's the problem with that?
<sphalerite_>
I've got a bridge configured via nixos config, which an LXC container attaches to. Whenever I switch the config in a way that modifies the bridge (e.g. updating iproute2), the container gets removed from the bridge and I have to add it back manually for the container to ahave internet access again. What's the best way to preserve connectivity for the container automatically?
<hyper_ch>
infinisil: my script doesnt have a daemon mode.... I am pondering how to cron my script in the modules
<infinisil>
hyper_ch: daemon mode means just fork into background, which is not what you want with a systemd service, so the nixos module for znapzend doesn't even use that
coot has quit [Quit: coot]
<infinisil>
ottidmes: what's your github nick?
<hyper_ch>
infinisil: I mean for my backup script... how to make it run regularly
<hyper_ch>
with znapzend having a daemon mode - the daemon does the job
<hyper_ch>
my script not having one... I need to cron somehow
<infinisil>
hyper_ch: this has nothing to do with daemon mode
<infinisil>
znapzend is a binary that runs continuously, it manages the timing itself
alex`` has quit [Ping timeout: 252 seconds]
<hyper_ch>
isn't that what you call a daemon mode?
<infinisil>
daemon mode is commonly refered to forking the process into background
<ArdaXi[m]>
Specifically, disowning it so PID 1 becomes responsible for it
roberth has joined #nixos
<infinisil>
hyper_ch: but if you want your script to run at certain times, the easiest way is to use systemd timers
<hyper_ch>
infinisil: you can set that up hte mode module expression?
<infinisil>
or a cron job i guess, but I'm never using cron jobs, but i think they'll work too
<hyper_ch>
so, systemd now just needs to also integrate grub and the linux kernel
<infinisil>
huh?
<newbie__>
I tried to depend on unstable cmake in my nix-shell, and I got "addEnvHooks: command not found".
Unode_ has joined #nixos
<hyper_ch>
infinisil: (do one thing and do it well - systemd is way past that)
<ArdaXi[m]>
Systemd already had a bootloader, systemd-boot
<ArdaXi[m]>
*has, even
<hyper_ch>
the bootloader formerly known as gummiboot?
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] Infinisil opened pull request #35073: nixos/znapzend: fix when no previous zetup (master...fix/znapzend) https://git.io/vAW1j
NixOS_GitHub has left #nixos [#nixos]
<infinisil>
ottidmes: Here is my PR with the patch^^
<gchristensen>
lol infinisil
<ArdaXi[m]>
Though subsuming cron is one of the best fits for systemd, it already ties in with services etc
<infinisil>
gchristensen: huh?
<gchristensen>
`zetup`
<infinisil>
heh
<infinisil>
well i thought i'd go with znapzend's ztandard naming
<ottidmes>
gchristensen: That is actually how znapzend calls its setups, so it matches the termology used
<gchristensen>
ah
<infinisil>
there's even znapzendztatz
<hyper_ch>
it's like that "k" thing for kde apps
<Unode_>
hi all, when installing a package (chromium) system wide using environment.systemPackages as customizing it's attributes (enableWideVine = true, enablePepperFlash = true), is the standard procedure to create a derivation or is there a simpler syntax?
<gchristensen>
unfortunately that timeout is global
orivej has quit [Ping timeout: 248 seconds]
bluemonk12 has quit [Quit: Leaving]
<Dezgeg>
let's hope one day there's no need to build a vendored llvm during the rust build
<gchristensen>
indeed
<gchristensen>
can it compile in parallel yet?
<LnL>
I didn't have success trying to pull that apart last time
<samueldr>
is it possible to build it into its own derivation?
<LnL>
but it should be possible to pull out the vendored build in a separate drv
<rardiol[m]>
could the build be split? so both halves take less than 10h? Build the vendored llvm, then actual rustc?
<gchristensen>
rardiol[m]: are you doing aarch64 stuff?
<LnL>
yeah, that's what I was trying before
<Yaniel>
gchristensen: crates are compiled in parallel and 1.24 changed the default llvm codegen units to 16
<Yaniel>
also enabled incremental compilation by default
<LnL>
we already enable codegen units for the build
<gchristensen>
Yaniel: Nix sets parallelism to 1 iirc
cransom has quit [Quit: WeeChat 1.7]
<rardiol[m]>
gchristensen: no, I just got an email because I'm a maintainer for pcsxr, which got a propagated failure from there and took at look at the problem
<LnL>
but it's still mostly bottlenecked by a single core
<sphalerite_>
newbie__: just a guess, but maybe try adding pkgconfig to the shell
Guanin has quit [Remote host closed the connection]
cransom has quit [Client Quit]
<sphalerite_>
that often helps finding libraries
<newbie__>
let's try :)
Guanin has joined #nixos
<newbie__>
sphalerite_: no luck :/
<sphalerite_>
don't know then, sorry :(
alex`` has joined #nixos
<dgpratt>
trying nixos on a VBox vm, I decided to switch from 17.09 to unstable -- running "nixos-rebuild switch --upgrade" produces errors relating to "virtualbox.service", e.g. "virtualbox.service: Main process exited, code=exited, status=1/FAILURE"; does anyone have a guess as to the cause or a possible resolution to that error?
<sphalerite_>
dgpratt: anything interesting in the logs? journalctl -eu virtualbox
zarel has joined #nixos
<LnL>
did you reboot afterwards?
<newbie__>
sphalerite_: lol if I install the packages and cmake outside the nix-shell, it sees eigen but not armadillo ^^
va1entin has joined #nixos
<va1entin>
Hey, does anyone know the status of support for AppImages in NixOS?
<newbie__>
sphalerite_: never mind, now it works. Not sure if and what I changed ^^ Thanks!
phaebz has joined #nixos
<dgpratt>
sphalerite_: nothing beyond what I see in the output of the rebuild command -- it does mention something like "failed to connect to the guest property service"
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] vcunat pushed 3 new commits to gcc-7: https://git.io/vAWDA
<NixOS_GitHub>
nixpkgs/gcc-7 e96741c Vladimír Čunát: treewide: remove explicit overrides to gcc7...
<NixOS_GitHub>
nixpkgs/gcc-7 eea5539 Vladimír Čunát: libmemcached: -fpermissive to fixup build with gcc7
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/gcc-7 7d19d7e Vladimír Čunát: Revert "linux-*: build with gcc7, but allow overriding it"...
<sphalerite_>
dgpratt: but yeah probably what LnL said — major version jumps like that won't work ootb without a reboot
<dgpratt>
ah, ok -- how can I verify it worked after a reboot?
<sphalerite_>
kind of funny how I can build stuff for an ISP-supplied consumer router, and run it on it, even though it was never designed for it, but not for a device which is intended for developing IoT stuff
<sphalerite_>
no idea how to work out why though :/
yann-kaelig has joined #nixos
<Dezgeg>
what's the problem?
<sphalerite_>
Dezgeg: getting SIGILL when trying to run a program and don't think I have any helpful tools for debugging it
<sphalerite_>
(host is openwrt)
<sphalerite_>
although I have got the process running in gdbserver and my laptop's gdb attached now
va1entin has quit [Ping timeout: 240 seconds]
<Dezgeg>
maybe you have a toolchain producing wrong kind of mips code
<sphalerite_>
AFAICT the instructions are all nonsense
<sphalerite_>
yeah, that's my guess as well
<Dezgeg>
is the endianness wrong? :)
<wzy8L-B2[m]>
Is it possible to use '.config/nixpkgs/config.nix' or something similar to declaratively install local packages like what is normally done in 'environment.system packages'?
<sphalerite_>
Dezgeg: I think the endianness was wrong *before*, it would always give me "no such file or directory" on exec
<Dezgeg>
on a static binary?
<Dezgeg>
well anyway, I would pull a known-working binary from the existing device and compare readelf output to the non-working binaries
<sphalerite_>
Dezgeg: yeah I think you're right it doesn't have the C1 coprocessor. Now to work out how to tell binutils/gcc that (or do you *happen* to know that too? ☺)
<archaeron>
hi :) I read somewhere that not all installed programs should go into `/etc/nixos/configuration.nix` wich sounds good since you need root for that. But I don't get (and can't find in the manual) how one would install something for a single user.
<sphalerite_>
Per-user package management is awesome
sud0x3 has joined #nixos
<archaeron>
i mean nixos-rebuild needs root. or maybe I'm doing this wrong
<hyper_ch>
yes, it needs root
<sphalerite_>
archaeron: don't worry, you're not doing it wrong. hyper_ch just isn't a fan of per-user package management apparently
<archaeron>
okay :) thanks.
<hyper_ch>
sphalerite_: I fail to see the point of per-user package
<ertes>
sphalerite_: thanks!
<archaeron>
then I'll do the ad-hoc thing for now
<hyper_ch>
and if the guide says not everything should go into configurtion.nix I think that's wrong
<hyper_ch>
(it probably depends a lot on what else is written there regarding that)
<samueldr>
it's different settings for different use cases
<sphalerite_>
archaeron: the simplest way to do declarative user package management is using an overlay like LnL's https://gist.github.com/LnL7/570349866bb69467d0caf5cb175faa74 but there are other ways, another popular one is home-manager which gives you nixos-like configuration for your user environment
<samueldr>
on a single-user machine, it's perfectly right to put everything into configuration.nix
<archaeron>
fair enough. but on a multi.-user machine?
<samueldr>
but it's not the only way to do
yann-kaelig has quit []
<archaeron>
but I will do it like that for now. Thanks!
<LnL>
I think the main reason a lot of people don't like user packages is because the default interface is imperative
<hyper_ch>
what does it hurt if user A has access to k3b and user B doesn't?
<archaeron>
another question. how do I query nix-env for non free packages?
<sphalerite_>
Myrl-saki: the magic of user namespaces
<Myrl-saki>
sphalerite_: Oh mkay.
Theuni82 is now known as Theuni2
<Myrl-saki>
sphalerite_: I was also thinking of doing user namespaces with /nix, is that possible?
<sphalerite_>
that's pretty much what nix-user-chroot does
Theuni2 has quit [Client Quit]
<ertes>
halfway through eelco's talk, and i'm already installing nixUnstable =)
Theuni2 has joined #nixos
<Myrl-saki>
sphalerite_: `unshare(): Operation not permitted()` :(
Theuni2 has quit [Client Quit]
<sphalerite_>
Myrl-saki: yeah then getting /nix set up for you, or enabling user namespaces, or something like that which all requires admin, will be necessary
<sphalerite_>
Myrl-saki: or you can use an alternative store path which is barely tested and results in having to build the world
<Myrl-saki>
sphalerite_: Thanks.
<Myrl-saki>
Wait
<Myrl-saki>
I have one more thing to try. proots!
<samueldr>
I was going to suggest looking at them
<sphalerite_>
Oh yeah,some people have got that working too iirc
<samueldr>
it's been reported in the past to work
<sphalerite_>
best solution is a real multi-user nix installation though, because that's the only way to get sandboxed builds as well
<Myrl-saki>
I think I'll ask the sysadmin to install it.
<Myrl-saki>
Is there a debian package for it?
<samueldr>
yes and no, yes there is one, but when it's mentionned it's often talked down as "that package" which "isn't supported"
<samueldr>
if you give me a minute I'll find the URL where the .deb is built on hydra
rogue_koder has joined #nixos
<Myrl-saki>
samueldr: Thanks. I'll try to talk to the admin to install it.
<samueldr>
it will need some configuration to work as a multi-user setup afterwards
<hyper_ch>
so, what wrong with this systemCronJobs entry: "*/15 * * * * hyper /run/current-system/sw/bin/r2e run" - it just won't run... only when I login as user and manually run it
alex`` has quit [Ping timeout: 240 seconds]
<hyper_ch>
Profpatsch: halp :)
Arcaelyx has joined #nixos
chreekat has joined #nixos
<samueldr>
hyper_ch: a shot in the dark, services.cron.enable is set to true?
<hyper_ch>
samueldr: yes, the other crons wrok just fine
<samueldr>
does the generated cron file look right?
<Profpatsch>
hyper_ch: Just create a simple systemd service and put something in startAt
<hyper_ch>
a simple systed service is very complicated for me :)
ma272 has quit [Quit: WeeChat 2.0]
<Profpatsch>
Just look at the about 2000 examples in the nixpkgs repository. :)
<samueldr>
and anything in the journal for --unit=cron ?
<hyper_ch>
still haven't figured out how to properly set external usb to offline with systemd when powering down
ma27 has joined #nixos
chreekat has quit [Client Quit]
cransom has joined #nixos
<hyper_ch>
samueldr: nothing suspicious in there
<cransom>
hrm. my old setup for weechat and python modules became ineffective.
<sphalerite_>
cransom: when?
<sphalerite_>
cransom: I made some pretty radical changes to that but it was months ago
<cransom>
i was using extraBuildInputs to feed in extra cruft for scripts and when i moved to 1.9/2.0 today it was ignoring them. though my fix at the moment seems slightly better as i don't have to recompile weechat if i add a thing.
ThatDocsLady has quit [Remote host closed the connection]
<samueldr>
hyper_ch: I don't know how r2e / rss2email works, but could it be succesfully "doing nothing"? as in it wants something in its environment but can't get it under cron?
<samueldr>
I'm thinking something like environment variables, as it's well probable that the invocation isn't with a login shell, thus rules differ
<samueldr>
(this is all baseless speculation obviously)
<hyper_ch>
samueldr: no idea :(
goibhniu has joined #nixos
<hyper_ch>
maybe the problem is the sendmail
<samueldr>
running `env -i $(which r2e) run` you could at least test whether rss2email is fine without any/most environment variables
<hyper_ch>
good thinking :) I just gave the path to sendmail in the config /run/wrappers/bin/sendmail and wait till next run before I try your
<samueldr>
no need for $(which r2e) as PATH seems to be evaluated before clearing env
nflores has joined #nixos
<samueldr>
(though it's highly probable that rss2email will barf without any environment variables!)
<sphalerite_>
Any cross wizards here who know how, with a working cross setup, I can get my hands on a binutils that will run on my laptop and whose objdump will be able to disassemble binaries compiled for the target platform?
manjaro-mate has joined #nixos
<manjaro-mate>
I have a question about NixOS's future path. I read a blog that NixOS is phasing out SystemD. I hope this is true. Can anyone confirm or deny please?
<hyper_ch>
where did you read that?
<manjaro-mate>
A post on the Arch Linux forum.
<hyper_ch>
samueldr: that didn't help
<hyper_ch>
so what di you want me to run now?
<samueldr>
if I understood correctly, rss2email works when you run it manually under your user
<sphalerite_>
manjaro-mate: not really true. Some people are working on supporting other init systems, but the standard is planned to remain systemd to my knowledge
<hyper_ch>
samueldr: if I run manually, then it works
<samueldr>
running it the same way, but prefixing with `env -i` will clear the environment
<sphalerite_>
manjaro-mate: make that "not true at all"
<samueldr>
oh sure, it was only a long-shot to see whether rss2email would barf or quit silently when it can't run due to missing environment
<Neo-->
anyone using pypi2nix and had issues with osx runs? on nixos the package (gevent) builds nicely, however on osx it fails to build (https://gist.github.com/uskudnik/0bf38a1b0d21e7ec7cdaf3fa401aedf4). gevent==1.1.2 builds on OSX too, but I'm not sure how to best debug what changed to fix generation of python expression
<manjaro-mate>
sphalerite_: Thank you for the post. I've been on FreeBSD since systemd screwed my clients. Linus is easier and better supported. I look forward to the day when systemd goes away.
<manjaro-mate>
I'm currently running manjaro in a VM.
<manjaro-mate>
Thank You for the info.
<Neo-->
garbas, around by any chance?
manjaro-mate has quit [Quit: Leaving]
<sphalerite_>
manjaro-mate: there's this "suckless nixos" or SLNOS project going on
<sphalerite_>
but
<sphalerite_>
aw ffs
<srhb>
sphalerite: Does it have any kind of internet presence aside from that mailing list thread and the occasional systemd criticism in here?
zzamboni has joined #nixos
<srhb>
oxijs github maybe?
<srhb>
Hmm, doesn't look like it...
vcunat has quit [Read error: Connection reset by peer]
vcunat has joined #nixos
<nflores>
sphalerite_: i tried the curlOpts you mentioned last night but curl seems to complain with curl: (6) Could not resolve host: application, I wonder if we need to escape "Accept: application/octet-stream" in the curlOpts?
<hyper_ch>
samueldr: so, any other idea?
chreekat has joined #nixos
<hyper_ch>
from where is the user env read?
<sphalerite_>
nflores: oh no, shell escaping strikes again. Yeah, if you just stick lib.escapeShellArgs before the list it should hopefully work
<samueldr>
other than using something else to verify the crontab line runs successfully (like a script that date > ~hyper/some_file.txt) then no, sorry
chreekat has quit [Client Quit]
<nflores>
sphalerite_: thanks!
<samueldr>
all my issues with cronjobs, from experience, were stemming from wrong environments or assumptions of login shells vs. non-login shells
<sphalerite_>
srhb: and the withering opposition to my PR the other day… I'm pretty sure it does, but it's not public
<samueldr>
(and that was on non-nixos)
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
nixpkgs/master edc4a1f rnhmjoj: nixos/tests: fix random failures of privacy test
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 3df3ec8 Joachim F: Merge pull request #35079 from rnhmjoj/privacy...
<NixOS_GitHub>
[nixpkgs] joachifm pushed 2 new commits to master: https://git.io/vAWdR
<hyper_ch>
samueldr: let's see what it does with */10 * * * * hyper . /etc/profile; /run/current-system/sw/bin/r2e run
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] joachifm pushed 1 new commit to master: https://git.io/vAWdE
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/master 49b0760 Yuri Aisaka: cpp-gsl: init at unstable-2018-02-15 (#34998)
<samueldr>
hyper_ch: a more regular way to do this is to use something like `bash -l -c "whatever"` AFAIK, but login shells aren't really a discoverable feature :|
boegel has quit [Read error: Connection reset by peer]
NixOS_GitHub has joined #nixos
<Fare>
(In case I didn't say it here, my earlier problems where with -f '<nixpkgs>' vs -f $nixpkgs when NIX_PATH didn't use $nixpkgs, and somehow the two variants of gcc 6.4.0 were not compatible.
<NixOS_GitHub>
[nix] dezgeg opened pull request #1876: Minor interpreter performance improvements (master...perf-fixes) https://git.io/vAWFt
NixOS_GitHub has left #nixos [#nixos]
<Fare>
(from nixos-17.09 vs nixpkgs-unstable)
* sphalerite_
facepalm
vaibhavsagar_ has quit [Quit: ZNC 1.6.5 - http://znc.in]
vaibhavsagar_ has joined #nixos
<ertes>
i had an experiment going a few months ago to write an OS based on nix and nixpkgs that used a more algebraic approach to system configuration, and one of the design goals was "no systemd"
<ertes>
but that went nowhere mostly because i couldn't actually find an alternative
<sphalerite_>
what was missing?
<ertes>
sphalerite_: an alternative to systemd
<sphalerite_>
yes, but I mean from other init systems/service managers
<ertes>
sphalerite_: linux-specific features… i wanted cgroup/namespace management, but to my knowledge none of the services managers/process supervisors support that
<sphalerite_>
ah right
<sphalerite_>
not supported by ones that target other platforms because it's linux-specific and not supported by linux-specific ones because there's already systemd for that I'm guessing x)
<ertes>
yeah, that sounds about right =)
<ertes>
what i was looking for is a simple process supervisor with linux-specific features… i guess i could write one, but it's a lot of work
<dgpratt>
please forgive the noobish question, but I just installed firefox and I can run it from a terminal, but it doesn't show up in the gnome desktop, nor does it show up when I search; is there a simple way to make firefox known to the gnome shell?
<sphalerite_>
dgpratt: not sure if gnome-shell picks profile changes up automatically
<sphalerite_>
dgpratt: try alt-f2 and type r
<sphalerite_>
to restart the shell
<dgpratt>
sphalerite_: that did the trick
vaibhavsagar_ has quit [Quit: ZNC 1.6.5 - http://znc.in]
endformationage has joined #nixos
<dgpratt>
you're just a whirling dervish of helpfulness :)
<sphalerite_>
jtojnar: I think this sort of thing should be picked up automatically by gnome-shell, do you maybe have any idea about that?
<sphalerite_>
hahaha thank you
<duncan^>
ertes: Did you have a look at skarnet s6 as service manager?
zzamboni has quit [Quit: Leaving.]
Joppe has joined #nixos
cransom has quit [Quit: WeeChat 2.0]
<ertes>
duncan^: yes, i'm actually a heavy user of s6
<ertes>
but it's not good enough
cransom has joined #nixos
<duncan^>
What features do you need from systemd that you cannot find elsewhere?
<cransom>
sphalerite: thanks for the pointer. i would have caught docs sooner but master hadn't caught on yet.
<ertes>
duncan^: see above
noffle has joined #nixos
<duncan^>
ah, didn't see scrollback
<sphalerite_>
cransom: no worries, sorry for the breakage!
<noffle>
hello. trying to grok what the expr <nixpkgs> evaluates to in a nix expr I run with nix-build. does it look at my current nix env to resolve that?
<sphalerite_>
noffle: NIX_PATH
<cransom>
feature, not a bug. i don't mind bleeding from the cutting edge of unstable from time to time!
<ertes>
duncan^: in fact even with my current use cases i often run into the problem of runaway processes, because s6 only really cares about the processes it supervises
jrolfs has quit [Ping timeout: 268 seconds]
<ertes>
by its very philosophy it's not supposed to fix that though, so this is not just a bug that could be fixed
<ertes>
systemd would actually handle that the way i want… but the thought of replacing s6 by systemd makes my skin crawl =)
<noffle>
sphalerite_: is the syntax <foo> evaluating to "pull in the exprs at $foo"?
<duncan^>
Time for systemd-mini ? ;\
<noffle>
oh, rather it looks for "foo=X" in $NIX_PATH
<sphalerite_>
noffle: it desugars to (builtins.findFile "foo"), which in turn evaluates to the first location at which foo can be found in NIX_PATH
<noffle>
right -- okay cool
mrkgnao has quit [Ping timeout: 252 seconds]
sud0x3 has quit [Quit: WeeChat 2.0.1]
<sphalerite_>
noffle: yes. "pulling in the expression" is then achieved by calling import on that
<noffle>
in this case "foo" is an.. attribute?
<Joppe>
Hi, anyone knows what exactly is needed to compile with clang? A simple prog gives 'iostream' file not found while g++ compiles the same file fine. (I saw some similar bug report but I have tried both 17.09 and 18.03 now)
<noffle>
or just an arbitrary key from NIX_PATH
<noffle>
?
<sphalerite_>
noffle: the nix pills, https://nixos.org/nixos/nix-pills/ are an in-depth introduction to the nix language which you may find interesting
<sphalerite_>
noffle: it's a path in the filesystem, discovered through NIX_PATH
<sphalerite_>
noffle: an arbitrary one at that, but one that exists
<noffle>
sphalerite_: thanks; I'll look at that
nflores has quit [Ping timeout: 260 seconds]
tgunb has quit [Ping timeout: 240 seconds]
<srk>
looks like my hydra builds expect "builtin" platform/system due to change in nix.. had to add systems = [ "builtin" "x86_64-linux" ]; to localhost
jrolfs has joined #nixos
archaeron has quit [Read error: Connection reset by peer]
raynold has quit [Quit: Connection closed for inactivity]
shabius has quit [Quit: Leaving]
<srk>
win, hydra building with two netbooted slaves \o/ everything deployed with nixops
<srk>
unionfs benchmark
kim0 has quit [Quit: Connection closed for inactivity]
<srk>
heh, didn't last long invalid ownership on file ‘/nix/store/5j74c2agb6vrsrfwv520vwvhwlb4sdqf-perl-5.24.3/bin/perlbug’
tgunb has joined #nixos
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] StijnDW opened pull request #35081: brlaser: init from latest git (master...master) https://git.io/vAWN4
NixOS_GitHub has left #nixos [#nixos]
<srk>
looks like unionfs issue
smichel17 has quit [Quit: smichel17]
<hyper_ch>
is r-raymond here?
drewr has quit [Ping timeout: 276 seconds]
dejanr has joined #nixos
<srk>
hmm, would it be possible to get rid of the unionfs?
shabius has joined #nixos
ambro718 has joined #nixos
<ambro718>
I have a problem with deploying a machine with nixops, related to user/groups config. The users and groups I configured are not created. Here is deploy output: https://paste.ubuntu.com/p/b2SmSvd6NP/
<ambro718>
notice the "malformed JSON string" error
<clever>
ambro718: does it still fail at offset 856?
<ambro718>
it does
<clever>
"nixbld27"
<clever>
856 is in the middle of this string, from the list of groups
<clever>
next thing to check would be the store integrety
<clever>
ambro718: ssh into the router, and run nix-store --verify --check-contents
phdoerfler has quit [Ping timeout: 260 seconds]
jrolfs has joined #nixos
<ambro718>
it is running, it's not a very fast box
phdoerfler has joined #nixos
kelleyNif has joined #nixos
<phdoerfler>
Hi! How do I get cron to send me emails?
<dgpratt>
besides trying to mentally parse the expression, is there a better way to try to understand the difference between e.g. "gitMinimal" and "gitAndTools.gitFull"?
<clever>
dgpratt: gitFull includes tools like `git gui`, which depend on the X libs, causing it to take more disk space up
<srk>
oh yes, I've rebooted my hydra workers with overlayfsed nix store and it fixed the perl issue. nice
<dgpratt>
thanks clever
<srk>
overlayfs also seems to work much faster than unionfs-fuse
<clever>
error: attribute ‘sendmail’ missing, at (string):1:1
dejanr has joined #nixos
<clever>
phdoerfler: that package doesnt exist
<phdoerfler>
clever: oh
<clever>
srk: that should work without the fstab manipulation
<ambro718>
clever: no errors
<srk>
clever: will try to submit PR with what he suggests
codygman has joined #nixos
<srk>
clever: without any changes to stage-1?
<phdoerfler>
clever: the example value just had `/nix/store/.../bin/sendmail` in the nix file so I just guessed that ${pkgs.sendmail} would be the right thing
<clever>
ssmtp.out 0 s /nix/store/r5sidvlw2jmajbdzwbkgaqyq1cmc7vjf-ssmtp-2.64/bin/sendmail
<srk>
now there's a case for if [ "$fsType" = unionfs-fuse ];, I guess we can add enother if needed
<clever>
postfix.out 310,456 x /nix/store/pkix4gxx84py3197miaw9iw150nvjfz3-postfix-3.1.3/bin/sendmail
<clever>
exim.out 0 s /nix/store/9a3gxna4qgwkfkg1cg633glrgy4nh201-exim-4.89/bin/sendmail
<clever>
phdoerfler: 3 packages provide a sendmail binary, youll need to pick the one you like the most
dejanr has quit [Client Quit]
<phdoerfler>
I guess I have to pick the one that comes with the mail server I am using? OK that will do. thanks clever!
<codygman>
I'm trying to update nixos and it keeps failing on flash_player_npapi: http://lpaste.net/6930178311918190592 Any suggestions? I thought running sudo nix-channel --update && sudo nix-channel --list && sudo nixos-rebuild switch would be enough, but as seen by the log I linked it wasn't :/
davidak has quit [Remote host closed the connection]
<srk>
codygman: nixos-rebuild switch --upgrade
<srk>
clever: thanks
dejanr has quit [Client Quit]
davidak has joined #nixos
<clever>
srk: qemu_test1 will build the netboot image from <nixpkgs> and boot its kernel+initrd inside qemu with a dummy disk drive
<clever>
srk: qemu_test2 then boots the same disk drive, without the kernel+initrd
<Neo-->
garbas, if you have any ideas from the top of your head I have machine here to test it out
<srk>
cool
<clever>
so you can verify the installation works
<clever>
srk: this also adds the justdoit script to the installer, so you type justdoit, and its done
<phdoerfler>
!m clever
<[0__0]>
You're doing good work, clever!
<srk>
clever: like I can now make changes in nixpkgs and run nixops deploy to rebuild and redeploy all the images on netboot server but that is starting to take a while :D
shabius has quit [Quit: Leaving]
<clever>
srk: and you have to reboot real hardware, that lacks copy/paste while debugging
* srk
needs gigabit to rack
rogue_koder has joined #nixos
<clever>
but qemu just ignores the NIC entirely, and you can screenshot
<clever>
srk: curses is a weird one, because it doesnt use the serial port of the guest
<srk>
right
<clever>
srk: the VGA card in the guest has a block of ram that is filled with raw text, and the GPU is responsible for turning that into an image
<clever>
and it just doesnt turn it into an image, lol
<ottidmes>
Is there a way to fix this warning I get each build: install-info: warning: no info dir entry in `/nix/store/8gls8w2wijml91nf0v53r9bb0amkhfxb-system-path/share/info/automake-history.info'
<srk>
lol :D
davidak has quit [Remote host closed the connection]
<codygman>
Can I update my revision to c1d9aff56e0? I thought that's what nix-channel --list did
<srk>
was wondering how the hack..
<clever>
ottidmes: you could try programs.info.enable = false;
davidak has joined #nixos
robstr has quit [Ping timeout: 240 seconds]
<clever>
codygman: c1d9aff56e0 is what your on already, and thats too old
<clever>
codygman: what does sudo nix-channel --update say now?
<clever>
codygman: the fix is from 9 days ago, and is on the 17.09 branch of -channels...
<clever>
codygman: the tip of the 17.09 branch is from 4 days ago
<garbas>
Neo--: can you open an issue, i'm not sure i'm going to have time tonight to look at this
<Neo-->
garbas, sure!
dejanr has joined #nixos
<clever>
codygman: and your on a commit from 18 days ago, aha
shabius has joined #nixos
<clever>
codygman: and howoldis says that the channel changed 1 hour ago, to a commit from 18 days ago
<garbas>
Neo--: it looks like you should be adding additional build time dependency for gevent
<clever>
codygman: temporarily switching to the nixos-17.09-small channel should fix your problem
jb554 has quit [Ping timeout: 255 seconds]
<codygman>
clever: Huh, that's odd... but okay
<ambro718>
clever: maybe I could try loading the JSON file with perl?
<clever>
ambro718: maybe?
<ambro718>
but I don't know how to get the JSON module working, did "nix-shell -p perlPackages.JSON" but it still says Can't locate JSON.pm
<clever>
ambro718: read the perl script from the error, and also reuse the perl in its #!
<Neo-->
garbas, that was my hunch as well but various combinations that I tried didn't get it to work. on the other hand, on nixos there are no additional dependencies listed in requirements.nix
<fragamus>
does anyone have a nixos setup for haskell development with all the cool stuff - I am reinventing some wheel here and I would like to get the best dev setup for nixos and haskell
<garbas>
Neo--: did you look what buildInputs are defined in nixpkgs?
<clever>
srk: ah, that old error :D
<srk>
clever: what does even /etc/Fixups mean?
<clever>
srk: absolutely nothing
<clever>
srk: but nix doesnt know that, and blindly keeps those outputs around
<Neo-->
garbas, yup, tried with -E libev too but to no avail
<clever>
which prevents them from being GC'd
<clever>
so nix doesnt have to rebuild them later
<srk>
ok
<srk>
sounds like fun issue
<clever>
the problem, is that the /bin/sh in the sandbox, depends on the .lib for glibc
<clever>
so a glibc impurity is always added to the sandbox
<garbas>
Neo--: aha you are trying to run pypi2nix on osx.
<srk>
ah, right
<clever>
but if you are rebuilding the .doc for other reasons, the .lib is silently dropped
<clever>
leaving you with a /bin/sh that cant find its own libc
<srk>
lol
<clever>
and now system() fails
<Neo-->
garbas, correct :)
<garbas>
Neo--: i misunderstood you are only trying to run nix-build requirements.nix
<fragamus>
im looking for a spacemacs haskell setup also
<clever>
if the binary cache is enabled, it can just grab .doc from there
<clever>
but deepfire was running without any cache
<clever>
ambro718: we have been looking at the wrong json the entire time, lol
<znc-user>
test
<ambro718>
ah right
zzamboni has joined #nixos
<clever>
ambro718: that file is how nix remembers the semi-random uid mappings, so when you delete a user, then re-add it, it gets the old uid again
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] dotlambda opened pull request #35084: nixos/home-assistant: test MQTT configuration (master...home-assistant) https://git.io/vAWpA
NixOS_GitHub has left #nixos [#nixos]
<ambro718>
clever: that would be an empty file on the remote machine
<clever>
ambro718: it must have been corrupted
<clever>
ambro718: just delete it and see if that fixes things
<ambro718>
possibly by an interrupted deploy
zzamboni has quit [Read error: Connection reset by peer]
zzamboni1 has joined #nixos
<ambro718>
weird, I would expect it to use atomic replacement to prevent this problem
<Neo-->
garbas, possibly :) i'll try to take a look tomorrow again :)
<clever>
ambro718: ext4 puts more focus on the metadata, over the data
<clever>
ambro718: improper shutdowns can truncate files
<clever>
ambro718: zfs just rolls back the entire change, leading to the old version
<Dezgeg>
well, the perl script should be writing it in an atomic way
<clever>
Dezgeg: it is, it writes to a .tmp, then uses rename to move it
<ambro718>
clever: had to remove the gid file as well. Works now. Thank you.
<clever>
Dezgeg: but the file still wound up 0 bytes in size
<Dezgeg>
no, that's not crash-safe
<clever>
fsync between the 2?
<Dezgeg>
yes
Fare has quit [Ping timeout: 276 seconds]
<ottidmes>
lol, I wonder how my laptop with ext4 is still functioning then (same installation for like 7 years), since I forget to plug in the power plenty of times :P
<Dezgeg>
and after as well, I guess you don't want it to be rolled back on crash either
<clever>
ottidmes: its mostly about chance, was something writing at that moment
Joppe has quit [Ping timeout: 260 seconds]
<Dezgeg>
it's not only ext4 but pretty much any Linux filesystem
robstr_ has quit [Remote host closed the connection]
<ottidmes>
clever: I am then always greeted with those inode journal messages when I boot into the machine again, so it at least seems to notice on some level that things went wrong
<ottidmes>
I am now always using btrfs or ZFS or FAT32 (USB/EFI)
<sphalerite_>
ugh FAT32
<clever>
i wonder...
<sphalerite_>
number of times that has broken my boot by crashing before the kernel and initrd were written to the ESP…
<Dezgeg>
a good reason not to use EFI
<clever>
how complex would it be to write an EFI driver for ext4 or ZFS, stick it on the fat32, and chain-load a 2nd bootloader?
<clever>
in a way that obeys the efi specs and grants all efi programs usage of those FS's?
<Dezgeg>
I imagine you could reuse U-Boot
<ambro718>
I wrote an async FAT32 implementation in C++ some time ago. I have no idea how FAT32 file systems aren't crashing all the time or how one would implement it to reduce the chance without major performance loss or wear.
<ottidmes>
I just got a machine with EFI for the first time, I had my hopes up, but I actually prefer the BIOS interfaces
<Dezgeg>
ambro718: it's actually relatively possible for many operations, we sell such a product at $work
<sphalerite_>
yeah EFI isn't really that great. What I really want is coreboot everywhere
<clever>
sphalerite_: i had an idea yesterday, could a coreboot image that doesnt deal with dram/pci initialzation, be jammed into the boot rom of a network card
<sphalerite_>
yeah I saw that
<dgpratt>
I'd like to check a couple assumptions I have regarding how nix-shell and/or nix-build work, based on this: https://github.com/dmjio/miso/blob/master/sample-app/default.nix#L1 -- firstly, I think the "pkgs" parameter would normally be supplied (?), so the first bit that seems to be importing some nix expression directly from a git commit would not normally take effect?
<clever>
sphalerite_: then you tell the existing bios to network boot, and it hands control over to that boot rom
<clever>
sphalerite_: then coreboot finishes the job
<ambro718>
Dezgeg: that is good to hear. Still I imagine it takes 10x more code than a "simple" implementation.
<sphalerite_>
clever: it's somewhat nice in that it allows booting off non-EFI-supported filesystems, but a major gripe I have with OEM firmware is that it's really bloody slow
<clever>
sphalerite_: another benefit, is that you cant really brick the machine, since the original bios can just be told to not netboot anymore, then reflash the nic again
<clever>
but yeah, you still have to deal with the original POST
<sphalerite_>
like is it really so hard to initialise hardware that it takes 10 (or 180 for some servers) seconds before it can hand over to something the user controls
<srk>
clever: yeah, that etc/Fixup hack helped
<clever>
sphalerite_: yeah, that just makes no sense at all
<Dezgeg>
I hear some ECC ram needs every single byte written at boot, or something
<Dezgeg>
but yeah, even still obviously there is some extreme crap in server BIOSes to take that long
<WilliamHamilton>
hi, if there is someone here that has used minizinc in nixos, I'd like to know what the workflow is for executing `.mzn` files; I successfully used `mzn2fzn -> fzn-gecode`, but I wonder why there isn't a `minizinc` executable like it's hinted by tutorials and courses online. Pinging fpletz and infinisil for the github commits
<ottidmes>
My laptop exists the BIOS super fast, so fast that I sometimes miss pressing e.g. Esc on time when I want to enter the BIOS, but all my other machines take like 15-20s which is insane
<clever>
Dezgeg: ah yeah, the initial random state would have crc failures everywhere
<Dezgeg>
I presume the CPU is also running at low clocks at boot
<sphalerite_>
ottidmes: IMHO google got it right with the chromebooks. The firmware is so fast that you can't catch it. You can however tell it to wait for you by pressing some extra buttons along with the power button
<clever>
Dezgeg: this reminds me, one of my older motherboards, it must have a micro-controller on it somewhere
<sphalerite_>
(and it *just so* happens that that firmware is coreboot xD
dejanr has joined #nixos
<sphalerite_>
)
<clever>
Dezgeg: on bootup, it says things, in english, via the sound card, giving you progress updates
<clever>
Dezgeg: one day while messing with overclocking, it said, "no cpu found"
<sphalerite_>
whaaaaa
<Dezgeg>
yeah I remember such things as well
<clever>
that same bios, also lacks bounds checking on its reflashing utils
dejanr has quit [Client Quit]
* sphalerite_
searches youtube for talking bios
<clever>
you can record your own messages in windows, and reflash the bios with the .wav files
<clever>
but if you speak too slowly, it takes up too much space
<sphalerite_>
nnnope
<clever>
and the bios forgets how to boot from ide :P
<sphalerite_>
oh boy
dejanr has joined #nixos
<clever>
but it can still boot from cdrom
<clever>
and the win98 cd has an idiot mode, that chainloads the hdd, because you cant remove a silly cd
justan0theruser has joined #nixos
<clever>
in my case, it was an asus p4b motherboard
<clever>
that should help with the googles
justanotheruser has quit [Ping timeout: 255 seconds]
<clever>
i should just plug the thing back in and record it for historical purposes, lol
dj_goku has quit [Ping timeout: 260 seconds]
dejanr has quit [Client Quit]
dejanr has joined #nixos
<prooftechnique>
So, I know about using nix-build -A to get the store path for a package. Is there a way to make this work for overridden packages?
<clever>
prooftechnique: where is the override?
zzamboni1 has quit [Quit: Leaving.]
<prooftechnique>
In my packageOverrides
shabius has quit [Quit: Leaving]
<clever>
prooftechnique: config.nix overrodes?
<prooftechnique>
Yeah
<clever>
nix-build obeys that file automatically
dj_goku has joined #nixos
dj_goku has quit [Changing host]
dj_goku has joined #nixos
<prooftechnique>
Hmm. Weird. It doesn't seem to be picking it up.
<clever>
re-run nix-build with -v and check the output, you should see it reading config.nix
<prooftechnique>
(I'm using nix on Darwin with nix-darwin, which I assume is relevant)
<prooftechnique>
In particular, I'm trying to get nix-build to give me the store path for myrust
shabius has joined #nixos
<clever>
prooftechnique: thats not a config.nix file
<clever>
prooftechnique: nix-build doesnt understand that
fragamus has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever>
prooftechnique: a config.nix file starts with just { packageOverrides = ...
<clever>
~/.config/nixpkgs/config.nix
<clever>
prooftechnique: and this is where it goes
<prooftechnique>
Ah. Maybe I should move my overrides to such a file, then.
<clever>
you can also put nixpkgs.config = import ~/.config/nixpkgs/config.nix; in the old file
<clever>
to make it reuse the new config.nix body
<newbie__>
Did anyone ever get this when loading numpy in a nix-shell (python 2)? ImportError: dynamic module does not define module export function (PyInit_multiarray)
<clever>
srk: each gpu driver within qemu has its own variant of text_update
<ambro718>
I want that for a specific package with it is included in environment.systemPackages, the /share is also linked into /run/current-system/sw/share
NixOS_GitHub has joined #nixos
<NixOS_GitHub>
[nixpkgs] 7c6f434c pushed 1 new commit to master: https://git.io/vAleY
<NixOS_GitHub>
nixpkgs/master 0ac5af9 Michael Raskin: lispPackages.clwrapper: patching the second shebang, because source builtin is used
NixOS_GitHub has left #nixos [#nixos]
kim0 has joined #nixos
<ertes>
is suppose with nix 1.12 i still use nix-env for manipulating environments?
<LnL>
prooftechnique: nix-build '<darwin>' -A pkgs.myrust
<srk>
eh, very helpful hydra error :D {UNKNOWN}: invalid keys in declarative specification file at /nix/store/2bwy8rfpsfs024mybms3m1w8zafxrg85-hydra-perl-deps/lib/perl5/site_perl/5.24.3/Catalyst/Model/DBIC/Schema.pm line 526
pkill9 has quit [Ping timeout: 265 seconds]
jrolfs has joined #nixos
civodul has joined #nixos
<ambro718>
Just earlier I added a filesystem to configuration (fileSystems.) where the device did not exist, and this causes the system to not boot at all.
<ambro718>
Is this expected, and is there a way to proceed boot regardless?
<clever>
ambro718: you can set the noauto flag, that should make it happy
<ambro718>
clever: I want it to be mounted automatically
<ambro718>
but not cause boot to hang if there is a problem
<NixOS_GitHub>
[nixpkgs] vcunat pushed 1 new commit to gcc-7: https://git.io/vAlvV
NixOS_GitHub has left #nixos [#nixos]
<NixOS_GitHub>
nixpkgs/gcc-7 a84844d Vladimír Čunát: Merge branch 'master' into gcc-7...
zarel has quit [Ping timeout: 240 seconds]
zarel_ has joined #nixos
hiratara has quit [Ping timeout: 255 seconds]
<ambro718>
Nixops is connecting to the wrong IP for virtualbox deployment (of wrong interface), how to tell it the IP? I set deployment.targetHost but it has no effect. I also tried to change deploment.targetEnv to "none" but that gives me an error that it is "currently unsupported" .
<srk>
ambro718: yeah, how it works is it creates the instance and stores the privateIpv4 of it (libvirt backend does it as well)
<ambro718>
I think what happened is one time it deployed, it found other interfaces on the machine and for some reason decided it likes one of those more than the right one
<ambro718>
so how would I fix it?
<srk>
I don't know of a way to change it other than export/edit/import or update in sqlite
<srk>
or test key value pull request for nixops that can use json
<ambro718>
I can't even log into the machine, I don't know the password, and manual ssh gives me a password prompt for some reason
<srk>
there's no password
<srk>
only keys if you supplied ones
<ambro718>
nixops ssh worked list time, but manual ssh doesn't now
<srk>
if not keys are generated by nixops and only nixops ssh works then
<infinisil>
ambro718: nixops has keys on its own, you can see everything nixops has in its state with nixops export
<clever>
ambro718: the ssh private key is inside the sqlite database
<ambro718>
oh how very convenient
<infinisil>
and you can change that json, then nixops import again
<clever>
ambro718: its very convenient, until things break :P
<ambro718>
found it though
<clever>
one min
<ambro718>
already got it, privateKey something
<srk>
sucks you can't just do something like nixops modifynode privateIpv4 bla bla
<clever>
ambro718: this will dump the entire state as json, then use jq to index into the deployment uuid, resources, hostname, and then the private key for the none backend
zzamboni has quit [Quit: Leaving.]
zzamboni has joined #nixos
<ambro718>
it has \n inside which presumably is not ok
<clever>
it should be split into multiple lines, at the \n's
<clever>
the -r flag for jq does that for you
<srk>
these builders running from ram seems to be quite fast now, gcc in like 24minutes
<srk>
22
zzamboni has quit [Client Quit]
<ambro718>
clever: that worked, I was able to log in, thanks
<srk>
ram jumps to 80Gb, then to 50, 20.. when the build is done it's on 18
<srk>
will see how long till it's unable to build anything
<srk>
ah now, much less
shabius has quit [Quit: Leaving]
<srk>
it's /10
<ambro718>
I still have the problem that nixops can't deploy, how would one fix the IP?
<clever>
ambro718: if you export the json to a file, you can edit the problem IP, then re-import it
<clever>
ambro718: id also make a backup of the sqlite before you do that
zzamboni has joined #nixos
<infinisil>
note that import has the --include-keys argument
vcunat has quit [Ping timeout: 248 seconds]
<ambro718>
"cannot delete this deployment because it still has resources"
<ambro718>
but I need to delete it to import it
<clever>
it may need a delete --force?
<ambro718>
ok that worked
<ambro718>
ok got it to work again, thanks everyone
<ambro718>
I'm sure it will break again at the next deploy though
jb554 has joined #nixos
zzamboni has quit [Quit: Leaving.]
<bhipple[m]>
Is there a way to specify a binary cache on the cmdline, or does it have to be in nix.conf?
<clever>
bhipple[m]: only root can do that, via --option binary-caches foo
<clever>
--option can override most things in nix.conf, but some require root to have an effect
<bhipple[m]>
Makes sense. On a multiuser non-NixOS install, does the nix daemon have to be run as root?
<bhipple[m]>
or can it be run as a nixdaemon user in the nixbld group?
<clever>
technically, it just has to be ran as the user that has +w to /nix/store/
<bhipple[m]>
with the nix store owned by nixdaemon
<clever>
so if the store is locked down and you use nix-daemon to gain access, you have to run it as the user with +w
<bhipple[m]>
The daemon doesn't need to do any fancy setuid or mounting or other root-required cmds?
<clever>
that would probably make the security worse
<clever>
the daemon just has to be started as the right user, by the system service manager
hiratara has joined #nixos
<Dezgeg>
I think it still runs as root to do setuid() for example
<bhipple[m]>
Yeah I think I'm going to give it a try so I don't get skeptical looks from sysadmins when I run this as root everywhere
<bhipple[m]>
but just wanted to double-check that it isn't a doomed expedition from the start
<clever>
bhipple[m]: also, i think nix-daemon will need root, since it will try to setuid the worker children, which must be members of the nixbld group
<srk>
clever: yeah, it won't creating that two folders by itself
<srk>
clever: also there's no chroot option so mount spits
<srk>
overlay on /nix/store type overlay (rw,relatime,lowerdir=/mnt-root/nix/.ro-store,upperdir=/mnt-root/nix/.rw-store/store,workdir=/mnt-root/nix/.rw-store/work)
<srk>
overlay on /nix/store type overlay (ro,relatime,lowerdir=/mnt-root/nix/.ro-store,upperdir=/mnt-root/nix/.rw-store/store,workdir=/mnt-root/nix/.rw-store/work)
jrolfs has joined #nixos
<clever>
srk: your creating an upper directory, but its not used
<srk>
how come? mount says it is
<clever>
srk: i think upper, is what you would point upperdir= at
<clever>
and your using an existing tmpfs
<clever>
also, workdir= could point to another tmpfs defined in fileSystems, which eliminates the bash entirely
<srk>
ah, right, so it might be better to point both upper and work to separate tmpfs-es?
<Dezgeg>
well you need some bash anyway to add the /mnt-root prefix only in initrd
<srk>
or just upperdir=/mnt-root/nix/.rw-store ?
<clever>
that would leverage nixos creating the directories for you
<srk>
will try
<srk>
makes sense..
<clever>
Dezgeg: if you tell nixos that a tmpfs is mounted at /nix/store, it should add the /mnt-root on its own
<ottidmes>
clever: You suggested to just use sleep to implement a timer in initrd (I was thinking to complex, so thanks for that suggestion), do you perhaps know if it is a good idea to use /proc/sysrq-trigger to implement the shutdown? I was thinking to first use it to terminate all processes except init, then swapoff --all and umount --all, and then just make it shutdown, that should be safe enough right?
<srk>
sync ; echo b > /proc/sysrq-trigger
<srk>
:D
<clever>
ottidmes: yeah, once youve closed all open mounts and ran a sync, it should be safe to cut power
<ottidmes>
clever: Right, forgot to mention the sync, I was planning to do that as well `echo s > /proc/sysrq-trigger`
raynold has joined #nixos
<srk>
umount will do the sync
spear2 has joined #nixos
<ottidmes>
srk: And will umount --all also take care of swaps, I think it should, but not completely sure
<srk>
ottidmes: ooh, not sure. you can do swapoff -a to be sure
<ottidmes>
srk: Yeah, cannot hurt
MP2E has joined #nixos
<tertleeltret>
is nix os installed manually?
<tertleeltret>
or is there a Vmware appliance?
<clever>
[392990.152375] overlayfs: filesystem on '/home/clever/.daedalus/upper' not supported as upperdir
davidak has quit [Remote host closed the connection]
davidak has joined #nixos
shabius has joined #nixos
davidak has quit [Remote host closed the connection]
pie___ is now known as pie_
davidak has joined #nixos
lord| has quit [Quit: WeeChat 2.0.1]
lord| has joined #nixos
<newbie__>
I'm trying to compile against eigen, but I get an error upon include. The CFlags in the pkgconfig is correct (points to $prefix/include/eigen3), but In the enviroment variables the include is for $prefix/include. Is that normal?
lopsided98 has quit [Ping timeout: 255 seconds]
lopsided98 has joined #nixos
bas_ has quit [Ping timeout: 260 seconds]
jordanmackie has joined #nixos
jordanmackie has quit [Client Quit]
jordanmackie has joined #nixos
<jordanmackie>
Anyone able to help me debug nixos-install? In way over my head here...
dan_b has joined #nixos
<samueldr>
jordanmackie: ask away
<samueldr>
I may not be the one helping you, but once the question is out there, it's easier for others to help you :)
jrolfs has quit [Ping timeout: 260 seconds]
<infinisil>
Okay now this is confusing me
<infinisil>
I thought putting "${./.}" somewhere in a derivation would convert ./. to a /nix/store path (import it), that is indeed what happens
<infinisil>
But I also thought I could avoid that by doing "${toString ./.}", so it would put the actual /path/to/the/nix/file there, and not a /nix/store path
jrolfs has joined #nixos
<infinisil>
But that doesn't appear to work for what I'm doing
goibhniu has quit [Ping timeout: 240 seconds]
davidak has quit [Remote host closed the connection]
davidak has quit [Remote host closed the connection]
davidak has joined #nixos
newbie__ has quit [Quit: Page closed]
<tertleeltret>
ty ottidmes
ma271 has joined #nixos
<ottidmes>
infinisil: What do you get than? It seems to work consistently for me
<infinisil>
ottidmes: yeah testing it too now, hold on i almost got it tracked down
<clever>
infinisil: toString turns it into an absolute path, and the nix sandbox wont allow you to access it
<jordanmackie>
nixos-install keeps crashing with rsync errors when processing the iwlwifi stuff (poorly worded because I don't really know what's going on) - can anyone help?
ma27 has quit [Ping timeout: 256 seconds]
<infinisil>
clever: yea i know
<clever>
infinisil: what are you trying to do?
<tertleeltret>
i didnt realize there was GUI installer
<infinisil>
clever: It's writing the path to an init.el for emacs, so it find it in my config and edit it
tgunb has quit [Ping timeout: 256 seconds]
<tertleeltret>
but apremade appliance takes the cake
<clever>
infinisil: ah, then emacs should read that path at runtime
<infinisil>
yeah
simukis has quit [Ping timeout: 276 seconds]
tgunb has joined #nixos
<samueldr>
tertleeltret: no gui installer, the install is still done using the console, but it runs in a graphical session with some graphical tools (like a web browser) availabl
<jordanmackie>
samueldr: any idea why nixos-install would crash when processing iwlwifi?
<samueldr>
is it consistent?
<samueldr>
like, re-running nixos-install still "crashes" at the same point?
<jordanmackie>
Yep, every time
<samueldr>
and then, are you able to provide a more detailed log?
<infinisil>
ohhhhhhhhhhhhhhhhhh
<infinisil>
I gots it
<samueldr>
(a photo of your screen, while imperfect, is enough)
<jordanmackie>
Sure, hang on...
<infinisil>
I called my package with callPackage, where as the first argument (the path to the nix file) was generated by some function, which i ended up using like `callPackage "${./.}/${filename}" {}`
<clever>
infinisil: that turns it into a string, which may break some things
<infinisil>
so ./. was /nix/store/... already, so the ./. in the nix file is there too
<clever>
infinisil: i prefer to keep it as a path for as long as possible